svn commit: r282985 - in head/sys: arm/annapurna arm/annapurna/alpine arm/annapurna/alpine/hal arm/conf boot/fdt/dts/arm

[Eitan Adler]

On 18 May 2015 at 08:41, Mathieu Arnold <m at absolight.fr> wrote:
> +--On 18 mai 2015 10:52:29 -0400 Ed Maste <emaste at freebsd.org> wrote:
> | On 18 May 2015 at 10:11, Mathieu Arnold <m at absolight.fr> wrote:
> |>
> |> Mmmm, if this can be done only for base, and not for ports, sure, but
> |> ports need to be able to add patches with CRLF endings, because upstream
> |> software comes in all flavors, including CRLF files we need to be able
> |> to patch.
> |
> | Sure it could, we could configure it however we want.
> |
> | Note that we'll likely need to support CRLF in base anyway - one
> | obvious example is test cases for CRLF handling in various tools.  I'd
> | imagine we could add an SVN keyword like fbsd:crlf=yes, along the
> | lines of fbsd:nokeywords=yes.
> |
> | Do you know how common CRLF or partial CRLF files are in the ports
> | tree?  If it's only a handful that scheme could work there too.
>
> $ ag -l '\r' /usr/ports/|wc -l
>       95
>
> there's a bit more than a handful, but it could work, yes.  If you want to
> have a look at a typical one, /usr/ports/shells/ibsh/files/patch-Makefile
> :-)
>
> But I think Eitan was talking about Phabricator, in this case, it'd need to
> work there too.

The rule can be limited to the non-contrib portions of the source
repository.  If svn blocks the commit, and that's enough, that's okay,
but if it will help to have phabricator report these as well, I could
set that up.



-- 
Eitan Adler
Source, Ports, Doc committer
Bugmeister, Ports Security teams